silly power supply

Right. Winding inductors is a crazy diversion from getting things done. It appeals to people who don't need (or want) to get anything done.

Why would anybody want to make that themselves? We can buy a reel and poke it into the pick-and-place. 92 cents each.

Why?

On a sunny day (Sun, 25 Dec 2022 12:07:59 -0800) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:

So is endlessly posting to Usenet :-) In the same time you wind 2 coils.

You make crap stuff and sell it as the ultimate thing. So much wrong with your designs,,,

We weren't talking about winding inductors - which you can buy of the shelf - but transformers, which have more windings, so you mostly can't.

Rubbish. It appeals people who are more interested than you are in getting stuff done more or less right.

It has just got two identical coils. You can do step and step down, but only by a factor of two, and there aren't base/gate drive windings.

There are lots of applications where it solves your problem, which is why Eaton offers the product range, and many more where it doesn't.

The obvious answer is, to reel up a few hundred at a time, to sell to Highland for 92 cents each... fortunately for John Larkin, this answer IS good enough to suit his needs.

So they can blame Larkin for mistakes, rather than their own engineers.

I wonder how accurate are the zeners. I need 0.1% @12V.

is pretty typical. It's roughly +/-5%. If you need +/-0.1% you need a temperature stable voltage reference, and you usually need to trim the output voltage to get exactly 12V.

offers a part where Analog Devices has done the trimming for you. Sadly the highest voltage on offer is 5.000V +/-5mV.

There are quite a few others others.

"The Art of Electronics" has a couple of pages on the subject. Buying a tightly specified part is only the first step in giving the customer what they think they want.

As long as they have the same temperature coefficients, perhaps i can live with 1%.

I'll ask if they can supply parts at 10V9(R), 10V8(G) and 9V1(B). I use them as "Silly Volt Meter" in conjunction with RGB LEDs.

Alternatively, i can have someone pretest and presort them (Zeners and LEDs).

Customers don't need to make sense, they just need to pay invoices.

There was some nonsense about ionization at 15 volts. The photodiode is abs max 25 but the graphs show it working best at 75.

But might last least at higher voltage. You are literally burning it.

Taking the blame is OK but costs more.

0.1% without some sort of trimming, pot or dac, would be tough. Tempcos will getcha too.

People buy it.

Do many people buy yours? Full of hand-wound coils?

The figure of merit is

(externally perceived performance) / (internal simplicity)

and reliable too.

Yes, see my reply to Bill. Exactly voltage and/or color are not important, but i want them consistent within a group. I want R at 11.50V, G at 11.75V and B at 12.00V; so, color shift from Red to Yellow to White. When it's "White Hot", i need to attach 50W shunt regulators.

Not "fortunately" that I can make a power supply from available, inexpensive parts that we buy by the reel. That takes design.

Charge pumping off the stock 1:1 transformer saves making a custom part with multiple windings.

Wind your own transformers if you enjoy that. Strip and tin #30 enameled wire. Color stripe them to keep the ends right, or ohm them at assembly time. There's lots of amusement to be had from coil winding machines, solder pots, epoxy, thru-hole headers, and test sets.

I'd use a pretty good 5 volt reference and a trimpot and an opamp to make 12.00, and a few 0.1% resistors to scale down from there.

LM4040 family is good. Or a fancy ADR part.

Too many parts. I would rather pay for pre-sorting zeners and post-sorting zeners/LEDs. Just need to group the similar set.

Yikes. You'll wind up throwing away a lot of parts and a lot of labor.

12 volt zeners have terrible tempcos. You should have a constant-current source too.

Why sort LEDs? What circuit did you have in mind?

No, i can use group of them in a box. They just have to be similar in the group.

They should have close to same temp inside the box.

"Silly Volt Meter".

Replacing my nose and finger meter. I can smell the 5W zener burning and/or touching it. Just need to jack it up to 50W temporary.

So, get some TL431's, which have (as required) good aging and low temperature coefficients, and a good precision resistor for most of the 9.1V drop. A few resistor arrays, series connected, and some selection of taps can get the last two or three volts sorted into tiny steps. As long as you solder up the parts before doing the tap selection, the heat and stress of assembly won't be an error source.

Edison-style invention requires a good imagination and a lot of junk. The art is to imagine what specific bits of junk will solve each novel problem; it maybe isn't mass-produced zeners that solve this one.